Synopsis

Keywords: RF Arrays & Systems, RF Arrays & Systems

Transmit arrays are essential to mitigate B₁⁺ inhomogeneity at extremely high static magnetic field strengths. In addition, the transmit elements must be arranged in multiple rows to achieve whole-brain coverage at 11.7T. Furthermore, concurrent field monitoring probes offer the possibility to improve MR image quality by correcting subject-specific k-space trajectories and monitoring real-time dynamic field fluctuations. The placement of these probes must be carefully controlled to achieve proper probe conditioning as well as to preserve the RF performance of the transmit array. This work presents the numerical simulation and optimization of a 16-channel-transmit array with integrated field probes.

Introduction

A workflow combining 3D electromagnetic (EM) and RF pulse design (RFPD) simulations to achieve efficient excitation within SAR constraints for transmit array designs at 11.7T was developed¹. The proof-of-concept study included an 8-channel-transceiver array with a folded-end RF shield to minimise radiation loss¹. We have extended this approach to the development of a 16-channel-dual-row-transmit array, which achieves whole-brain coverage at 11.7T. To increase the signal-to-noise ratio (SNR), it will be combined with a 32-channel-receive array. Additionally, concurrently monitoring dynamic field fluctuations during MR image acquisition at 7T and beyond can improve image quality and reduce artifacts². However, optimal placement of field probes is essential to achieve probe conditioning and minimal interaction with the transmit array.
This work presents the optimization of a 16-channel-dual-row-transmit array using the workflow above. Then,16-field probes are integrated into the numerical model of the optimal transmit array. The probe positions were controlled to preserve the array's transmit performance.

Methods

A 16-channel-transmit array (array-δ) was developed using the optimisation process presented earlier^1. To achieve whole-brain coverage, a dual-row configuration consisting of two rows of eight rectangular loops covering 180mm in the z-direction was selected. EM simulations were conducted in CST Studio Suite 2021 (Dassault Systems, France)³ to optimise the transmit performance while the optimum configuration for mitigation of RF field inhomogeneity was chosen by RFPD simulations⁴.
The numerical model included the transmit array, coil housing, scanner bore, discrete component losses, and cable losses up to the coil plug. The array elements were arranged on a fiberglass tube (ε_r=5.5, tanδ=0.04, inner diameter 285mm, wall thickness 2.5mm), and the lower row elements were rotated by 22.5° (Fig. 1a). The coil was loaded with Duke body model (virtual family, 2mm resolution) truncated below the level of shoulders. Each array element consisted of 11-fixed capacitors and one variable capacitor The capacitors were connected by 2mm diameter wires. Adjacent elements were decoupled using geometric overlap and the diagonal elements were decoupled using transformer decoupling⁵.
To reduce radiation loss at 11.7T, a folded-end shield¹ was designed and optimised in conjunction with the array. Previously, we found improvement in transmit performance by increasing the shield diameter¹, but this would offset the coil from the iso-centre. Therefore, in addition to 400mm (δ₁) and 450mm (δ₂) shield diameter, an asymmetric shield with a flat base (δ₃) was also investigated because the coil sits lower on the patient table while the sides can be extended up to 450mm in diameter (Fig. 1b). The shield length was varied from 260mm to 300mm in 10mm increments.
Fig. 1c shows the CAD model for the full apparatus consisting of the transmit array with a receive array and optimal probes’ positions. The field monitoring probes were placed in pockets created on the enclosure of the receive array. Since the probes are shielded, they were modelled as solid copper structures in CST. However, their coaxial cables were not included. Positions for the probes were determined iteratively using a MATLAB programme (Skope MRT AG, Switzerland) and CST to achieve optimum probe conditioning and to minimise their influence on transmit efficiency.

Results and Discussion

Simulated results in Fig. 2 show that the array with an asymmetric shield (δ₃) electromagnetically performs equally as array δ₂, which has 450mm shield diameter. The asymmetric shield was chosen for further optimisation due to its lower height. Remarkably, this configuration also exhibits smaller NRMSEs of the k_T-points for all three different pulses (Fig. 2c). Similar to the previous study¹, we found no direct relationship between the shield length and radiated power. Among the arrays under investigation, δ₃(280) was preferred because it provided the least radiation loss, a compact design, and a more homogeneous B₁⁺ field (Fig. 2d and 2f). Fig. 3 presents the simulated S-parameters of δ₃(280) loaded with the Duke model. At 499.415MHz, the average couplings between the adjacent, next-neighbouring, and diagonal elements are lower than -15dB with a maximum coupling of about -12dB. The reflection coefficients were better than -40 dB with no resonant peak splitting.
Using the software for probe conditioning, the following results were observed: for a FOV of 20cm at the coil’s iso-centre, the simulated mean field noise was comparable to the probe phase noise with a maximum value of fewer than three times the probe phase noise within 50kHz bandwidth.
Fig. 4 compares the CP mode B1+ maps for δ₃ (280) in three different cases: (i) without the probes, (ii) with the probes but without retuning, and (iii) with the probes and retuning. These results demonstrate that the transmit performance can potentially be preserved by the appropriate selection of the positions of the probes regardless of channel retuning. The routing of probes’ feeding cables as well as the placement of traps on the cables will be determined empirically.

Conclusion

We have presented simulation results of a 16-channel-dual-row-transmit array with commercial field probes for B₀-field monitoring for MRI at 11.7T. The array is currently under construction and will be combined with a 32-channel-receive array to increase SNR. Combining EM and RFPD simulations is a rigorous numerical method for coil design at 11.7T. Also, placing the probes in optimal locations can prevent RF interactions and hence, preserve the transmit efficiency.

Acknowledgements

This project has received funding from the European Union’s Horizon 2020 research and innovation programme under grant agreement No. 885876 (AROMA). The authors acknowledge great customer support and useful discussions with Simon Gross and David Brunner from Skope MRT AG.